Abstract
Objective To describe characteristics of published research on the safety and efficacy of vitamin K antagonists (VKA) for pregnant patients with antiphospholipid antibodies (aPL), including their methodological characteristics and knowledge gaps.
Methods This study followed the Joanna Briggs Institute methodology for scoping reviews and used the Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews protocol system. Studies were primarily identified through searching electronic databases including MEDLINE, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials. Study characteristics and outcomes were reported and described using customized charting tables.
Results Of 1528 publications, 17 remained in the final analysis. These reported up to 190 VKA-treated aPL-positive pregnancies diagnosed as antiphospholipid syndrome (APS); pregnancy cases were likely overlapping in some publications. In the 17 reports, there were 723 individuals in comparison groups, including healthy pretreatment pregnancies and women with APS treated with standard therapies without VKA. However, only 4 (23.5%) of the 17 publications stated a study objective focusing on VKA use, of which only one was a full-length article. In addition, information on VKA doses, disease diagnostic criteria, and the long-term outcomes of offspring were largely absent.
Conclusion The current evidence is insufficient to assess VKA efficacy and safety profiles in aPL-positive pregnant patients. Studies with a defined focus on VKA use in this population are lacking, and reporting of key information is not consistent. The relative lack of knowledge of VKA use in pregnant women with APS is concerning, and efficacy and safety questions remain.
Antiphospholipid syndrome (APS) is a thrombotic autoimmune condition that is associated with adverse pregnancy outcomes. APS diagnostic criteria continue to evolve1 but invariably require clinical history of thrombosis (thrombotic APS) and/or pregnancy loss (obstetrical APS, including loss due to stillbirth, recurrent miscarriage, and placental insufficiency) with laboratory evidence of antiphospholipid antibodies (aPL). APS treatment during pregnancy commonly involves low-dose acetylsalicylic acid (ASA) with or without subcutaneous administration of low molecular weight heparin (LMWH) and oral hydroxychloroquine, resulting in a live-birth rate of more than 70%.2 Although many patients with APS have good outcomes with this treatment, there is a small subset of patients who experience adverse pregnancy outcomes, including placental insufficiency and pregnancy loss.3-5 We also note that some studies observed inadequacies of the standard ASA and LMWH treatment to prevent a recurrent cerebral ischemia in APS pregnancies with a history of cerebrovascular events.6,7
Vitamin K antagonists (VKAs), including warfarin, are used to prevent a recurrence of thrombotic events in patients with APS.8,9 However, VKA exposure between 6 and 9 weeks of gestation carries a risk of embryonic toxicity, including spontaneous abortions and congenital anomalies.10-12 In non-APS conditions, relative risk of birth defects may be 2- to 4-fold compared to nonexposed controls.11 The dose threshold of warfarin embryopathy may exist at 5 mg/day.13,14 In addition to embryopathy, second and third trimester exposure is reportedly associated with the fetopathy, including central nervous system abnormalities,10 cerebral bleeding, microcephaly, and optic nerve atresia.11,15 Although a dose dependence of fetopathy is not well defined,16 > 5 mg daily dose may be considered a threshold.13,17 Due to these risks, VKA treatment as secondary thromboprophylaxis in patients with APS is switched to low-dose ASA and subcutaneous LMWH once pregnancy is confirmed.18 At present, the use of oral VKA in pregnant women with APS after 12 weeks’ gestation of the postembryopathy critical period is reserved for exceptional circumstances.4,19 Although what constitutes the exceptional circumstances is specific to patient care settings and is difficult to categorically define, a history of arterial thrombosis may be one of the indications.6,7
In patients with APS with a history of venous thromboses, the target international normalized ratio (INR) is 2.0-3.0,20 which is lower than in a mechanical heart valve.21 The guidelines for achieving a target INR 2.0-3.0 in nonpregnant patients are to start VKA doses between 5-10 mg/day for the first 2 to 3 days, although a 5-mg starting dose may be preferred.17,22 In contrast, a target INR may need to be higher in patients with APS with a history of arterial thrombosis (INR > 3.0), but the consensus is not clear.23 Further, anticoagulation is also important in preventing fetal injury in APS-complicated pregnancies.24-26
Although most patients with APS are effectively treated with LMWH and low-dose ASA throughout pregnancy, VKA use after the first trimester remains a realistic option for a small subset of the patients, including those with a history of arterial thrombosis.6,7 As a precursor to knowledge synthesis for guiding the management of such women, we conducted a scoping review of research evidence of VKA use for pregnant patients with APS, identifying the types of research, geographical variations, and knowledge gaps. The temporal profile of these publications over the last 30 years has been also described. Throughout this scoping review, we use the term APS to describe the study populations unless stated otherwise.
METHODS
Protocol. The study was conducted using Joanna Briggs Institute methodology for scoping reviews and followed Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR)27 2009 guidelines (Supplementary Table S1, available with the online version of this article), as the protocol28 and initiation of the study was done before updated PRISMA-ScR guidelines were published.
Search strategy. Searches were first conducted in MEDLINE, Embase, Web of Science, and the Cochrane Central Register of Controlled Trials in January 2021, with search results updated up to February 2024. Both database subject heading fields (eg, MeSH in MEDLINE, Emtree in Embase) and text word fields were searched for the concepts of VKA, pregnancy (including maternal exposure, congenital abnormalities, teratogenicity, and embryo and fetal development), and aPL. Synonymous terms were first combined with the Boolean “OR.” These 3 concepts were combined with the Boolean “AND.” In all databases, truncation symbols and adjacency operators were used in text word searches when appropriate to capture variations in spelling and phrasing. No language or date restrictions were applied (for complete Embase search strategy, see Supplementary Table S2, available with the online version of this article). In addition, we reviewed citations of publications as an additional search for appropriate articles. When information was inadequate, we attempted to contact the article authors for clarification and details.
Citation management. All citations were imported into the bibliographic software manager Zotero (Corporation for Digital Scholarship). Duplicates were removed automatically and then manually when needed. A PRISMA flow chart of publication screening is illustrated in Figure 1.
PRISMA flow diagram. Adapted from Moher et al.53 For more information, visit www.prisma-statement.org. APL: antiphospholipid antibody; PRISMA: Preferred Reporting Items for Systematic reviews and Meta-Analyses.
Study selection. We examined all publications that reported pregnant patients with defined presence of aPL and their maternal and fetal outcomes. We included publications that describe such women with VKA exposure confirmed by noting gestational timing during pregnancy. Although our original plan was to not include publications with duplicated patient populations, we were unable to clearly define the magnitude of overlapping. Therefore, an emphasis was placed on delineating the characteristics of published evidence, and we included publications describing apparently overlapping groups of VKA-exposed pregnant patients. When this was the case, however, we added notes for clarification.
Common regimens with VKA during pregnancy are based on limiting its use to the second trimester to reduce risk of the warfarin embryopathy, with a critical exposure period between 6 and 12 weeks of gestation. Our focus was on studies describing those exposed to VKA after pregnancy confirmation, but not on preconceptional VKA use. Preconceptional VKA use is discontinued and replaced by LMWH-based therapy before or as soon as pregnancy is confirmed, unless medically indicated otherwise. Accordingly, we excluded reports if they describe only such VKA discontinuation cases unless VKA-related adverse outcomes were reported. We also excluded reports of animal studies and those without original data.
To assess the type of study included, we used the conventional classification of study designs,29-31 which defines case reports and case series as retrospective descriptions of ≥ 1 patients without an internal control group. However, they may include historical or pretreatment outcomes as a control.29 In contrast, a cohort study is characterized by exposure-based sampling of a population of interest with a prospective follow-up over time. These studies may or may not include a control group without the exposure of interest that is similarly followed to the outcome.30,32
Data extraction and charting. Two independent reviewers (JZ and JL) carried out the title and abstract screening and data extraction for included studies. Disagreements were resolved by discussion and consensus. Authors were contacted regarding incomplete data in the original article.
Data from eligible publications were charted using a customized extraction form. This form captured relevant information on key study characteristics and details on information used to describe pregnancy outcomes. Briefly, the following data items were charted: country of origin, types of study designs (case report, case series, cohort study with or without a comparison group), outcomes reported, author names, language, journal, and year of publication. In the reported outcome table, the following variables were extracted: APS criteria, the number of patients allotted to the intervention and comparison groups, specific treatment dosage, duration, INR, and outcome measures (maternal mortality and thrombosis, live births, embryopathy, pregnancy complications, fetal bleeding, small for gestational age, preterm birth, fetal loss).
RESULTS
Database search. The initial and updated searches identified 1528 results. After deleting 256 duplicates, an additional 1035 articles were excluded because they did not meet inclusion criteria based on their titles and abstracts. Reasons for nonselection were animal studies, reviews, and publications that did not include cases of APS or VKA as treatment. Two-hundred thirty-seven full articles were reviewed against eligibility criteria and 17 articles remained for inclusion in the scoping review (Figure 1). Basic characteristics of these studies are seen in Table 1.
Basic study information of 17 publications.
Included studies. These 17 publications reported a total of 190 VKA-treated APS pregnancies with some overlap in the patient populations (Table 1).6,7,33-47 Of the 17 publications, 10 were descriptive studies. In the remaining 7 reports with comparators, 723 comparison pregnancies were either healthy women without APS, the studied patient population in a previous untreated pregnancy, or female patients with APS treated with non-VKA standard treatment at the time. This non-VKA therapy included low-dose ASA with/without heparin, LMWH, or prednisone.
The following study types were noted (Table 1 and Figure 2): (1) a case report (5/17 [29%])33-37; (2) a case series of VKA use in pregnancy (2/17 [12%])38,39; (3) a case series of APS pregnancy that includes VKA-exposed pregnant women as a subset (5/17 [29%])40-43,46; (4) a cohort study of VKA use in APS pregnancy (2/17 [12%])44,45; and (5) a cohort study of APS pregnancy that includes VKA-exposed pregnant women as a subset (3/17 [18%]).6,7,47
Study type and publication format. Five study types exist, including 5 case reports, 2 VKA case series, and 2 VKA cohort studies. The remaining 8 publications describe VKA-exposed patients as subsets within their respective populations of patients with APS. The publication formats vary widely. APS: antiphospholipid syndrome; VKA: vitamin K antagonist.
Study objectives, design, and publication format. There were only 4 publications of systematic research with VKA treatment focus during pregnancy (4/17 [23.5%]; Table 1, Figure 2, and Figure 3): 2 VKA cohort studies with internal comparison groups44,45 and another 2 VKA case series.38,39 Of the 2 VKA cohort studies, the report by Jesus et al44 described the outcomes of 29 VKA-exposed APS pregnancies in comparison with low-risk pregnancies, but it was a conference abstract with few details. The other by Pauzner et al45 was an original research paper comparing the warfarin and the nonwarfarin-treated APS pregnancies. They concluded a favorable profile of efficacy and safety of VKA therapy. Of the 2 VKA case series, the one by Hunt et al38 was a correspondence, describing 2 pregnant women with APS with exposures at the first and second trimester or the second trimester only. The other, by Levy et al,39 described pregnancy outcomes of 79 APS pregnancies exposed to VKA in the second trimester, which was a short section of a narrative review. We identified 5 case reports of APS pregnancy33-37 describing VKA exposure during pregnancy, including 1 patient who used it against medical advice.36
Schematic representation of characteristics of studies on use of VKA antagonists in pregnant women with APS. * 2 studies7,35 reported confirmed35 and suspected7 warfarin embryopathy. The latter study is a case series of 23 preconception use of warfarin in pregnancy with APS, which was included in the present review due to the potential fetal adverse outcome. APS: antiphospholipid syndrome; VKA: vitamin K antagonist.
In the other cohort studies and case series of APS pregnancy without a particular focus on VKA use, there were 8 publications (3 APS cohort studies6,7,47 and 5 APS case series40-43,46) that included a subset of VKA-exposed pregnant patients. Note that 1 of the above 3 APS cohort studies described suspected warfarin embryopathy within an APS cohort of preconceptional VKA use followed by non-VKA regimen during pregnancy7; we included this publication in the category of APS cohort with VKA-exposed subset because the authors comment on potential warfarin embryopathy in 1 case. Of these 8 publications, 2 case series described subsets of 19 and 28 VKA-exposed pregnancies,42,43 but pregnancy outcomes were shown as a whole. The remaining 5 reports each described a subset of only 2-3 individual patients.6,40,41,46,47
Study location and publication year. These 17 reports varied in their study locations (Table 1 and Figure 3). Except for 1 case report from Canada,36 there was no systematic investigation on this topic in a North American setting. Among those with VKA research focus, a cohort study44 and a case series39 were from the same center in Brazil, a cohort study45 was from Israel, and a case series was from the United Kingdom.38
Four reports of systematic investigation with focus on VKA treatment (2 cohort44,45 and 2 case series38,39) were published between 1998 and 2010 (Figure 3). The most recent case series of 43 pregnancies,43 in which 28 cases were exposed to VKA, was published in 2014. Since then, only 2 case reports were published on the use of VKA in pregnant women with APS.36,37 The lack of VKA-focused investigation, particularly in the recent years, coincides with the current views against the use of VKA for most patients with APS during pregnancy.4,5
Overlap of study subjects. Studies from the same treatment centers may include overlapping cases. Although we were able to confirm case overlap in studies from Brazil,39,43,44 it was not possible to determine the exact magnitude of case overlap in part due to evolving APS criteria (Figure 3; G. de Jesús, MD, email communication, February 2019). However, we included these publications to best delineate published information, rather than to estimate the risk of adverse outcomes. Three women receiving VKA-based therapy described in each of the 2 articles from another group6,47 are seemingly overlapping, but it could not be confirmed. We included both publications because of their separate study foci. We also suspect some levels of case overlapping among studies from the same institution in Israel,40-42,45 but we could not confirm nor rule out the possibility.
Mode of VKA use. Eight reports (8/17 [47%]) depicted preemptive use of VKA for prophylactic purposes in the second or second and early third trimester39-42,43,44,45 or throughout pregnancy34 (Table 1). Another 8 (47%) described its use in response to a triggering event, including vascular events,6,33,35,38,47 heparin-induced thrombocytopenia,37 intolerance to LMWH,46 and against medical advice.36 In 4 of these 8 reports, the exposure started in the first trimester.33,35-37 As described above, 1 publication of suspected warfarin embryopathy was an APS cohort study of preconceptional VKA use.7
VKA therapy–specific outcomes. All 4 publications of a systematic study with a stated VKA-specific objective38,39,44,45 reported pregnancy outcomes of the VKA therapy. On the other hand, those without a stated objective of a VKA focus6,42,43,46 reported outcomes as a group of all treatment regimens that included VKA-exposed subsets, in which VKA-specific outcomes were not described. These articles reported that adverse maternal outcomes (preterm delivery, preeclampsia, bilateral notching on arterial Doppler) and fetal outcomes (fetal deaths, intrauterine growth restriction [IUGR]) were more likely to be associated with a history of thrombotic APS6,42,43,46 and the presence of lupus anticoagulant.6,42
Reported information. Although all publications reported live birth numbers, information on the presence or absence of thrombosis and adverse pregnancy outcomes, as well as details on health of the neonate exposed to VKA in utero and their long-term health, were often missing. Similarly, key information regarding doses of VKA and patient diagnostic criteria (eg, aPL types) were not always present (Table 2).
Reported information on APS diagnosis and VKA regimen (N = 17 studies).
VKA safety. Many of the publications did not explicitly report the presence or absence of VKA embryopathy/fetopathy, although it seems unlikely that they would not have described fetal abnormalities if they existed. We identified 1 case of VKA embryopathy35 and another case of nasal hypoplasia as suspected VKA embryopathy.7 In the case report,35 a patient with arterial occlusion was treated with phenprocoumon during the preconception and postconception periods until week 10 (2 weeks after the pregnancy confirmation) and delivered a newborn with coumarin embryopathy. In the case series of APS pregnancy,7 a woman was on warfarin preconceptionally until the time of pregnancy confirmation at 5 weeks’ gestation, at which point warfarin was switched to LMWH. The infant was born at 34 weeks due to fetal distress with IUGR and marked nasal hypoplasia, one of the signs of warfarin embryopathy. In addition, there were 5 cases of VKA exposure in the first trimester (1 of the 2 in a case series38 and 4 in case reports33,34,36,37), but embryopathy was not observed in these 5 cases (note that 1 of the 5 cases underwent therapeutic abortion at 12 weeks due to the complicated maternal condition37). The reasons for the use of a VKA in the first trimester of these 5 cases were indication for arterial thrombosis,33 unrecognized pregnancy,38 preemptive use for a history of heparin intolerance,34 heparin-induced thrombocytopenia,37 and against medical advice.36
Maternal complications of these patients included 1 patient with arterial thrombus at 25 weeks. In this case report,36 the patient chose to remain on warfarin against medical advice as she had experienced 6 pregnancy losses (embryonic and fetal) and attributed her pregnancy losses to the lack of efficacy of non-VKA–based treatment. Additional maternal complications associated with the first trimester use of VKA included 2 cases of severe preeclampsia at 31 weeks,33,38 and spontaneous preterm delivery at 35 weeks.34
Institutional ethics approval. A prospective cohort study by Stone et al47 with a focus on endothelial cell activation in APS pregnancy, which included a subset of 3 pregnant women on VKA, is the only publication that documented its institutional ethics approval. An additional study46 stated approval by the hospital’s audit committee. Notably, a clearly documented institutional ethics approval was not found in any of the 4 publications with a stated focus on VKA regimen in APS pregnancy (a conference abstract,44 a letter to the editor,38 a full article in 2001,45 and a narrative review with a section describing a case series39). This may be due to their abbreviated publication formats.
DISCUSSION
In this scoping review, we identified 17 publications addressing VKA use during pregnancy in patients with aPL,6,7,33-47 but the evidence gaps in the literature were striking, as briefly described below, which precludes systematic review approaches.48 First, there are only 4 publications with clear descriptions of VKA therapy in APS pregnancy,38,39,44,45 in which only 1 cohort study45 was a full-length article. The remaining 3 were a cohort study in a form of conference abstract,44 a case series of 2 patients,38 and a case series within a section of a narrative review.39 Although their study designs may be considered relatively high in the hierarchy of evidence level, the publication formats often prevent detailed analyses. Second, there was a lack of consistency in reported items (Table 2). Important missing items include APS disease criteria used for the study population, aPL profiles, detailed doses and durations of VKA use, pregnancy outcomes, and long-term outcomes of offspring. Considering the established dose and time-dependent embryopathy caused by warfarin, it is unfortunate that these details are often missing.
We found apparent geographical variations in the published practices of VKA use (Figure 3). Warfarin was used preemptively due to lack of accessibility of LMWH in Brazil around the time of their publications.39,43,44 The increase in accessibility has since shifted their practice to include LMWH preference compared to warfarin (G. de Jesús, MD, email communication, February 2019). In contrast, groups in Israel conducted an observational study on different treatment options based on stratifying baseline risks.45 The only publication in North America was a case report of a woman who continued warfarin against medical advice.36 Elsewhere, warfarin was used on seemingly a case-by-case basis when standard treatment options were insufficient or not tolerated.6,38,40-42,46,47
The center where warfarin replaced LMWH at 14 weeks, as part of a predefined treatment protocol regardless of risk factors,39,43,44 did not observe an increase in adverse obstetrical or thrombotic events compared to the non-VKA comparison groups. Similar outcomes were reported by Pauzner et al45 and Langevitz et al41 when warfarin was used preemptively during the second and part of the third trimester of pregnancy in women with higher risk of obstetrical and thrombotic complications,45 as well as in those with an increased risk of stroke.41 In these studies, pregnancies were not associated with fetal warfarin toxicity. Of note, Pauzner et al45 was the only publication we could find of a prospective study describing head-to-head comparison between the preemptive use of VKA in the second trimester and the non-VKA–based standard treatment in patients with APS. Since the publication by Pauzner in 2001,45 there has been no comparative cohort study on the use of VKA in pregnant women with APS, and their findings remain to be replicated. We are aware that a randomized controlled trial based in Mansoura University in Egypt is currently underway to compare second trimester warfarin use to LMWH use in pregnant patients with APS.49 This study was initiated in 2013 and is actively recruiting 100 participants with a goal of completion in 2024.
One of the limitations of this scoping review arose from the study selection strategy: publications were not included unless VKA exposure timing in pregnancy (indicating gestational weeks and/or trimester) and outcomes were noted. Although it is important to specify exposure timing in pregnancy for teratogenic drugs like VKA, some publications with unspecified gestational timing of VKA exposures were not analyzed. Similarly, we excluded publications that addressed pregnancy outcomes specific to preconceptional VKA use in women with APS, although we assessed 1 study7 with a suspected case of warfarin embryopathy. In general, pregnancy is confirmed around 5 to 6 weeks’ gestation on average.50 Therefore, preconceptional VKA use is associated with early pregnancy exposure even if it is replaced by heparin formulations as soon as pregnancy is confirmed. Although preconceptional use of VKA poses a question of warfarin embryopathy that is different from the fetopathy risk associated with the midpregnancy use of VKA, studies of preconceptional use of VKA would have been informative to further delineate the evidence gap in this field. Second, due to the substantial but unspecified levels of patient overlap among some publications, we could not follow our original study protocol to eliminate them. Although this constrained our analyses, we were able to highlight the inherent limitations stemming from the characteristics of the evidence in this field. Third, because of the lack of long-term follow-up of the exposed offspring, the presence or absence of VKA fetopathy remains unclear in these analyzed reports. Finally, the current scoping review highlighted the challenges for future systematic evidence synthesis, showing that reported studies were variable in their study scopes, designs, and outcome reporting. Also, studied patient populations were not homogeneous due to the evolving disease criteria, which may contribute to differences in treatment responses. In addition, the type of aPL was often not described in the publication, making it impossible to characterize aPL type–specific outcomes of VKA treatment, although the history of thrombosis was the most significant factor for future thrombosis/obstetrical complication (regardless of aPL profile). These details of the disease subtypes will be important for future studies.
In conclusion, the current prevailing view and the updated practice guidelines do not recommend VKA use for women with APS throughout pregnancy. However, a small number of patients, particularly those with a history of arterial thrombosis, may require more robust anticoagulation after 12 weeks’ gestation.6,7 The published evidence remains sporadic and insufficient to evaluate efficacy and safety of the midpregnancy use of VKA in those with APS, and this remains an open question. The need for studies on VKA use in thromboembolism prophylaxis with a specific focus on pregnant patients has been repeatedly mentioned.51,52 To this end, involvement of patient perspectives in conducting clinical research on drugs, including potentially embryotoxic and fetotoxic VKA, is important.
ACKNOWLEDGMENT
We express our sincere gratitude to Cheri Nickel and Glyneva Bradley-Ridout for their guidance throughout the literature review process. This work is based on part of the thesis of Judith Zarek, which was presented to the University of Toronto in partial fulfillment of the requirement of Doctor of Philosophy in Pharmaceutical Sciences.
Footnotes
DH was supported by the King Abdullah Scholarship, a program by the Saudi Ministry of Education.
The authors declare no conflicts of interest relevant to this article.
- Accepted for publication July 16, 2024.
- Copyright © 2024 by the Journal of Rheumatology
